coachwei

Like the result from the article. Showing the page load time vs percent of users is a very useful way to analyze site performance:how many percentage of site users are experiencing slow performance?

Further, combing this with web timing API would be even more interesting.

Great news. Google Chrome and FireFox both are adding support for Web Timing API. This is great for web development community (posted updated accordingly).

Good point. Tumblr is a lot faster than Posterous. However, it is only at about 50 percentile comparing to other sites over the Internet while Posterous is below average - in other words, Tumblr's performance is only about average. Both have lots to be improved.

just wrote about Tumblr for its web site performance (http://news.ycombinator.com/item?id=1712065). The technical team at Tumblr clearly did a good job. Sad to see CTO go.

Does a fast site help improve business success(comparing to a slower site similar in functionality)? Of course it does. Does performance play a role? It does. Is it the only reason? No.

The article does not pinpoint the success or failure of a business into a single reason. The article didn't claim "performance is The reason" that Tumblr is successful. It says "performance is most likely one of the key factors".

no, actually it does matter. Browser caching is not reliable and only covers around 20% to 25% of requests (according to GoogleMaps JS caching behavior stats). So it matters quite a bit.

Does web performance matter? For Tumblr and other light blogging service providers, clearly it does and it made a big difference. A recent New York Times article pointed out that "Tumblr Leaves Posterous In the Dust". Both are leading light blogging service providers. However, Tumblr has out-grown Posterous significantly over the last year. Why? it looks like that one big reason is that Tumblr has been twice as fast as Posterous.

The score shows as "0" initially until we've collected data from different geographic locations, which may take a little while Sometimes users mistake this "0" as the final score and didn't know they should wait until the score gets updated. We are aware of this issue and will fix this in future releases.

The average YSlow score for all sites is 69. Wah. Higher than I expected!

The article's author probably mistook "gigabit" as "gigabytes". Nevertheless, 40Gbps is impressive.